Report missing safe outputs in failure issue instead of failing conclusion job

actions/setup/js/handle_agent_failure.cjs

@@ -239,6 +239,7 @@ async function linkSubIssue(parentNodeId, subIssueNodeId, parentNumber, subIssue
 /**
  * Handle agent job failure by creating or updating a failure tracking issue
  * This script is called from the conclusion job when the agent job has failed
+ * or when the agent succeeded but produced no safe outputs
  */
 async function main() {
   try {
@@ -260,9 +261,21 @@ async function main() {
     // Check if there are assignment errors (regardless of agent job status)
     const hasAssignmentErrors = parseInt(assignmentErrorCount, 10) > 0;
 
-    // Only proceed if the agent job actually failed OR there are assignment errors
-    if (agentConclusion !== "failure" && !hasAssignmentErrors) {
-      core.info(`Agent job did not fail and no assignment errors (conclusion: ${agentConclusion}), skipping failure handling`);
+    // Check if agent succeeded but produced no safe outputs
+    let hasMissingSafeOutputs = false;
+    if (agentConclusion === "success") {
+      const { loadAgentOutput } = require("./load_agent_output.cjs");
+      const agentOutputResult = loadAgentOutput();
+
+      if (!agentOutputResult.success || !agentOutputResult.items || agentOutputResult.items.length === 0) {
+        hasMissingSafeOutputs = true;
+        core.info("Agent succeeded but produced no safe outputs");
+      }
+    }
+
+    // Only proceed if the agent job actually failed OR there are assignment errors OR missing safe outputs
+    if (agentConclusion !== "failure" && !hasAssignmentErrors && !hasMissingSafeOutputs) {
+      core.info(`Agent job did not fail and no assignment errors and has safe outputs (conclusion: ${agentConclusion}), skipping failure handling`);
       return;
     }
 
@@ -329,6 +342,15 @@ async function main() {
           assignmentErrorsContext += "\n";
         }
 
+        // Build missing safe outputs context
+        let missingSafeOutputsContext = "";
+        if (hasMissingSafeOutputs) {
+          missingSafeOutputsContext = "\n**⚠️ No Safe Outputs Generated**: The agent job succeeded but did not produce any safe outputs. This typically indicates:\n";
+          missingSafeOutputsContext += "- The safe output server failed to run\n";
+          missingSafeOutputsContext += "- The prompt failed to generate any meaningful result\n";
+          missingSafeOutputsContext += "- The agent should have called `noop` to explicitly indicate no action was taken\n\n";
+        }
+
         // Create template context
         const templateContext = {
           run_url: runUrl,
@@ -340,6 +362,7 @@ async function main() {
           secret_verification_context:
             secretVerificationResult === "failed" ? "\n**⚠️ Secret Verification Failed**: The workflow's secret validation step failed. Please check that the required secrets are configured in your repository settings.\n" : "",
           assignment_errors_context: assignmentErrorsContext,
+          missing_safe_outputs_context: missingSafeOutputsContext,
         };
 
         // Render the comment template
@@ -394,6 +417,15 @@ async function main() {
           assignmentErrorsContext += "\n";
         }
 
+        // Build missing safe outputs context
+        let missingSafeOutputsContext = "";
+        if (hasMissingSafeOutputs) {
+          missingSafeOutputsContext = "\n**⚠️ No Safe Outputs Generated**: The agent job succeeded but did not produce any safe outputs. This typically indicates:\n";
+          missingSafeOutputsContext += "- The safe output server failed to run\n";
+          missingSafeOutputsContext += "- The prompt failed to generate any meaningful result\n";
+          missingSafeOutputsContext += "- The agent should have called `noop` to explicitly indicate no action was taken\n\n";
+        }
+
         // Create template context with sanitized workflow name
         const templateContext = {
           workflow_name: sanitizedWorkflowName,
@@ -405,6 +437,7 @@ async function main() {
           secret_verification_context:
             secretVerificationResult === "failed" ? "\n**⚠️ Secret Verification Failed**: The workflow's secret validation step failed. Please check that the required secrets are configured in your repository settings.\n" : "",
           assignment_errors_context: assignmentErrorsContext,
+          missing_safe_outputs_context: missingSafeOutputsContext,
         };
 
         // Render the issue template

actions/setup/js/handle_agent_failure.test.cjs

@@ -26,7 +26,7 @@ describe("handle_agent_failure.cjs", () => {
 **Branch:** {branch}  
 **Run URL:** {run_url}{pull_request_info}
 
-{secret_verification_context}
+{secret_verification_context}{assignment_errors_context}{missing_safe_outputs_context}
 
 ### Action Required
 
@@ -40,7 +40,7 @@ When prompted, instruct the agent to debug this workflow failure.`;
       } else if (filePath.includes("agent_failure_comment.md")) {
         return `Agent job [{run_id}]({run_url}) failed.
 
-{secret_verification_context}`;
+{secret_verification_context}{assignment_errors_context}{missing_safe_outputs_context}`;
       }
       return originalReadFileSync.call(fs, filePath, encoding);
     });
@@ -550,14 +550,55 @@ When prompted, instruct the agent to debug this workflow failure.`;
   });
 
   describe("when agent job did not fail", () => {
-    it("should skip processing when agent conclusion is success", async () => {
+    it("should skip processing when agent conclusion is success with safe outputs", async () => {
       process.env.GH_AW_AGENT_CONCLUSION = "success";
 
+      // Set up agent output file with safe outputs
+      const tempFilePath = "/tmp/test_agent_output_success.json";
+      fs.writeFileSync(
+        tempFilePath,
+        JSON.stringify({
+          items: [{ type: "noop", message: "No action taken" }],
+        })
+      );
+      process.env.GH_AW_AGENT_OUTPUT = tempFilePath;
+
+      try {
+        await main();
+
+        expect(mockCore.info).toHaveBeenCalledWith(expect.stringContaining("Agent job did not fail"));
+        expect(mockGithub.rest.search.issuesAndPullRequests).not.toHaveBeenCalled();
+        expect(mockGithub.rest.issues.create).not.toHaveBeenCalled();
+      } finally {
+        // Clean up
+        if (fs.existsSync(tempFilePath)) {
+          fs.unlinkSync(tempFilePath);
+        }
+      }
+    });
+
+    it("should process when agent conclusion is success but no safe outputs", async () => {
+      process.env.GH_AW_AGENT_CONCLUSION = "success";
+      // Don't set up GH_AW_AGENT_OUTPUT to simulate missing safe outputs
+
+      // Mock API responses
+      mockGithub.rest.search.issuesAndPullRequests.mockResolvedValue({
+        data: { total_count: 0, items: [] },
+      });
+
+      mockGithub.rest.issues.create.mockResolvedValue({
+        data: {
+          number: 1,
+          html_url: "https://github.com/test-owner/test-repo/issues/1",
+          node_id: "test-node-id",
+        },
+      });
+
       await main();
 
-      expect(mockCore.info).toHaveBeenCalledWith(expect.stringContaining("Agent job did not fail"));
-      expect(mockGithub.rest.search.issuesAndPullRequests).not.toHaveBeenCalled();
-      expect(mockGithub.rest.issues.create).not.toHaveBeenCalled();
+      expect(mockCore.info).toHaveBeenCalledWith(expect.stringContaining("Agent succeeded but produced no safe outputs"));
+      expect(mockGithub.rest.search.issuesAndPullRequests).toHaveBeenCalled();
+      expect(mockGithub.rest.issues.create).toHaveBeenCalled();
     });
 
     it("should skip processing when agent conclusion is cancelled", async () => {

actions/setup/md/agent_failure_comment.md

@@ -1,3 +1,3 @@
 Agent job [{run_id}]({run_url}) failed.
 
-{secret_verification_context}{assignment_errors_context}
+{secret_verification_context}{assignment_errors_context}{missing_safe_outputs_context}

actions/setup/md/agent_failure_issue.md

@@ -4,7 +4,7 @@
 **Branch:** {branch}  
 **Run URL:** {run_url}{pull_request_info}
 
-{secret_verification_context}{assignment_errors_context}
+{secret_verification_context}{assignment_errors_context}{missing_safe_outputs_context}
 
 ### Action Required